I try to realize what is benefit of using Kafka stream in my business model. The customers publish an order and instantly gets offers from sellers who are online and intrested in this order.
In this case the streams are fit to join available sellers(online) to order stream and filter, sorting (by price) of offers. So as result the customer should give the best offers by price by request.
I discovered only one benefit: it is less of server calls(all calculations happends in stream).
My question is, why streams are matter in this case? Because I can make these business steps using the standard approach with the one monolithic application?
I know this question is opinion based, but after reading some books about stream processing it is still to hard change the mind on this approach.
only one benefit: it is less of server calls
Kafka Streams can still do "server calls", especially when using Interactive Queries with an RPC layer. Fetching data from a remote table, such as KSQLdb, is also a "server call".
This is not the only benefit. Have you tried to write a join between topics using regular consumer API? Or a filter/map in less than 2 lines of code (outside the config setup)?
can make these business steps using the standard approach with the one monolithic application?
A Streams topology can still be emdedded within a monolith, so I don't understand your point here. I assume you mean a fully synchronous application with a traditional database + API layer?
The books you say you've read should go over most benefits of stream processing, but you might want to checkout "Kafka Streams in Action" to specifically get the advantages of that